PDA

View Full Version : CHG_K8: New ADMIXTURE test



Fire Haired14
03-12-15, 14:32
CHG K8 Spreadsheet (https://docs.google.com/spreadsheets/d/1_uKagAzyBnSzfI1rIEclIx6kfIE-NYpiWY5Bs0XqyAg/edit#gid=1521430867&vpid=A1)
CHG K8 by Region (https://docs.google.com/spreadsheets/d/1JELnEcj_Q8_AKtFK_4z4UxBSn37QLnvrpe8yTbmN_Jo/edit#gid=20013436&vpid=A10)

CHG_K8 is the newest ADMIXTURE test made by Daviski to estimate ancient ancestry. It represents the most up to date knowledge on European and to some extent West Asian genetics. What we've learned is Europeans are a mixture of 4 distinct populations in West Asia/Europe from 8,0000+ years ago who were distinct because of 1,000s of years of isolation. And this test IMO gives pretty accurate percentages of ancestry from each group.

Some here have miss interpreted and miss judged these ADMIXTURE tests made by Davidski. There's nothing to suggest he's using these tests for an agenda. He isn't even claiming they are actual percentages of ancestry from each ancient population. It's just a suggestion and trend according to him. ADMIXTURE method isn't perfect and no one ever claimed it is.

Summary of Information it Reveals.

New components absorb old components.
>EHG component can explain the high ANE scores in the Caucasus, Siberia, South Asia, and Americas.
>CHG can explain most ANE in West Asia.
>Roughly half of the WHG Europeans scored in ANE K8 is absorbed by EHG/Anatolia Neolithic. Actual percentage of ancestry from Western HGs like Loschbour isn't above 30% anywhere according to this test.

SouthEast Europe and Caucasus are mostly CHG+EEF
>Modern Turkey and Caucasus are mostly CHG.
>South Caucasus/Turkey is CHG+EEF. North Caucasus is CHG+EEF+EHG.
>Greece-area, Italy, and Iberia are mostly European Neolithic.
>Greece-area and Italy have a lot of CHG.
>There isn't much EHG and therefore Steppe in Greece-area and Italy. According to this test 10-20%(if you dive Yamnaya EHG by their EHG). CHG-related pop could be the main newcomers after Neolithic in these two regions.

North Europe is LNBA+CHG
>North Europe has significantly less EHG than LNBA North Europe. And has CHG unexplained by Steppe. Bronze age trade with South Europe?
>Balkans look like a mix of North Europe and Greece.

SouthWest Europe is CHG+EEF+LNBA
>Sardinians and Iberians have significant CHG.
>There's 20% Yamnaya in Iberia, but if you use a LNBA North European as source of Steppe, population turnover in Iberia is almost 50%. Looks like a lot of CHG also came into Iberia after Neolithic.

Twilight
03-12-15, 18:55
Interesting, thank you for posting :)
How long ago do you think these ghost populations lived? It appears that the abbreviation dictionary has not caught up to words like EHG and CHG yet

Fire Haired14
03-12-15, 19:31
Interesting, thank you for posting :)
How long ago do you think these ghost populations lived? It appears that the abbreviation dictionary has not caught up to words like EHG and CHG yet

We have prove WHG and CHG were around 13,000 years ago. I imagine EHG was to. But, EEF may be a mixture of two populations that were around 13,000 years ago.

Angela
03-12-15, 22:41
Exactly as I've been arguing for two years or more, against much opposition I might add, and based solely on Dienekes admixture runs, the ancient dna papers as they started coming in, ancient history and logic.

One could look at it as two ancient "northern" components (WHG and EHG), and two "southern" components(Anatolian Neolithic and Caucasus). I said that southern Europe would wind up being 70-90% "Southern", and 30-10 "Northern". What does this "calculator" say? Many parts of Spain are 72% Southern (+ 2% Afro-Eurasian) and 25% Northern, Northern Italians are 75% Southern, The southern French are 70% Southern, Tuscans and Albanians and Central Greeks are 80% Southern, the Slovenians are 62% Southern, the Poles 54% Southern and on and on.

You could get the rough parameters from the original Lazaridis analysis, where the Anatolian Neolithic people and the CHG were obviously grouped together. It could also be calculated from the Haak et al three way split of Europe.

In some ways, you could get close to these percentages by using a few of the Dodecad calculators.

This analysis, if it stands up, raises some real questions about the estimates that have previously been proposed for the amount of actual Yamnaya ancestry in Europe, defined as a 50/50 split of EHG and CHG. It seemed to be the consensus on some sites that not only northern Europeans in general but also central Europeans are a 50/50 mix of MN and Yamnaya.I even saw people claiming the MN were almost totally wiped out and "replaced" by Yamnaya people. I don't see how that's possible if the percentages in this calculator are correct. The only people who might still work out as 50% Yamnaya might be the Balts or the Finns.*

This is an issue I've raised previously, saying that perhaps the group that went into central and northern Europe was more EHG shifted than Yamnaya, perhaps from the "Indo-Europeanization" of EHG forest steppe groups. However, the numbers here for a lot of even central European groups show less EHG and more WHG, and also more CHG. So, were there Indo-Europeanized groups which were more WHG shifted than the steppe people? ** However, what's the source of the additional CHG? It's explainable in various ways for southern Europe, but what about central and northwestern Europe?

Is it possible that the western Yamnaya region harbored a mix of WHG and EHG, or more WHG like EHG than those further east on the steppe, and somehow also harbored some CHG like population? I don't think that's impossible. When the J2 was found in Hungary I speculated as to whether a more Caucasus like population was starting to move into Europe at that time (mid to late Neolithic). I'd love to see the Kumtepe late Neolithic sample run through this calculator. The fact that the Greek sample from 4000 BC was very EEF like rather argues against that, but that might have been an anomalous result. Plus, I'd like to see that run through this calculator too. Or, the CHG might have begun to flow into Europe with the Copper Age. We have no idea of the autosomal composition of the CT people who moved onto the western steppe. Could they have mixed with a more western Yamnaya group?

In terms of the area of most interest to me, southern Europe, the results are also very interesting. As I've been saying ad nauseam, the "Caucasus" or "West Asian" in Italy can't be analyzed in a vacuum as something that only affected it and not neighboring countries. As I suspected, the CHG like levels are very similar in Greece, the Balkans, and Italy, and not all that much lower in Spain.

Central Greece- 45%, and that's after what, 25-30% "Slavic" admixture? What was it before? (Northern Greece is still 44%) Their total for ANeo and CHG is about 80%

Tuscans and Albanians-40-41%

Bulgarians 39%

Northern Italians-35%

Slovenians 34%

Poles and some Spaniards both have around 30-31%

So, when did it arrive in Europe? Late Neolithic? Copper Age? Bronze Age? Or were there repeated flows?

As to the ANeolithic figures, the highest are, of course, in Sardinia, then southern France, then Spain, then Bergamo and Tuscany, and then Albania, and then dropping from there.

Missing, of course, are the figures for southern Italy and Sicily, which for some inexplicable reason were not provided.

Ed.

* In their case, there was probably also a reservoir of WHG or WHG/EHG hunters who became Indo-Europeanized.

** Perhaps there were also movements around the same time from WHG groups who had been living at the periphery of northern Europe and in the general turmoil and climate change, started to filter southward.

One other thing that I thought I should mention in regard to CHG levels in southern Europe. All of the scenarios I mentioned above are possible, but there may be another factor as well. As I've speculated in the past, the "Indo-European" groups which went into the Balkans and Italy might have been CHG heavy, perhaps groups from the southern steppe. I've even proposed that it's possible that some of these "Indo-European" groups, specifically from the Yamnaya area from which we have samples, went around the Black Sea, through northern Anatolia and into Europe from that direction having picked up more CHG along the way, and so were more than 50% CHG by the time that they arrived. After all, the Yamnaya samples we have carried a form of R1b which is the most common R1b in southern Italy and Greece and the Balkans. Only samples from Anatolia and Greece from that period will tell us what happened.

Fire Haired14
04-12-15, 03:14
@Everyone,

I've added CHG_K8 to my 4mix spreadsheets. I can't think of all the differnt tests to do. So, please post ideas about tests to do.

You can learn how 4mix works here.

http://eurogenes.blogspot.com/2015/05/4mix-four-way-mixture-modeling-in-r.html

I've already found North Europeans only fit well as Steppe+EEF+WHG+CHG. Nothing else works. Here are examples of good fits.

Hungary: 31% Yamnaya_Samara + 39% Sweden_MN + 9% WHG + 21% CHG @ D = 0.0081
Norway: 37% Yamnaya_Samara + 36% Sweden_MN + 12% WHG + 15% CHG @ D = 0.0039
BeloRussia: 38% Yamnaya_Samara + 30% Sweden_MN + 15% WHG + 17% CHG @ D = 0.0119

Potapovka is a better source for Steppe ancestry in Europe so...
Hungary: 46% Potapovka + 20% CHG + 26% Sweden_MN + 8% WHG_CHG @ D = 0.0094
Norway: 56% Potapovka + 14% CHG + 20% Sweden_MN + 10% WHG_CHG @ D = 0.0059
BeloRussia: 59% Potapovka + 15% CHG + 13% Sweden_MN + 13% WHG_CHG @ D = 0.0126

Fire Haired14
04-12-15, 03:25
SouthEast Europeans and West Asians can take the place of CHG. When SouthEast Europeans are used Euro_MN disappears. An admixture with WHGs and SE Euro or West Asian in North Europe doesn't seem likely to me. If we take CHG_K8 literally there's significant SE European or West Asian ancestry in North Europe. One way or another extra CHG beyond Steppe got in North Europe.

Norway: 49% Potapovka + 35% GreekCHG + 4.99999999999999% Sweden_MN + 11% WHG_CHG @ D = 0.0043
Norway: 47% Potapovka + 39% AlbanianCHG + 3% Sweden_MN + 11% WHG_CHG @ D = 0.0042
Norway: 47% Potapovka + 43% Italian_BergamoCHG + 0% Sweden_MN + 10% WHG_CHG @ D = 0.017
Norway: 55% Potapovka + 17% GeorgianCHG + 17% Sweden_MN + 11% WHG_CHG @ D = 0.0049
Norway: 50% Potapovka + 27% Turkish_IstanbulCHG + 12% Sweden_MN + 11% WHG_CHG @ D = 0.02

Edit: If we take CHG K8 literally. It reveals three important things about Europe.
>There's West Asian CHG in North Europe unexplained by Steppe.
>There's little Steppe(based on EHG) in SouthEast Europe. Lots of West Asian CHG came in after Neolithic.
>South Europe is mostly Neolithic Euro. WHile nearby Turkey isn't. Turkey_Istanbul scores 23% Anatolia_Neolithic while Greek scores 35%.

Fire Haired14
04-12-15, 04:13
These fits are the best I can get for Tuscany and Greece. CHG_K8 suggests Greece area and Italy are mostly CHG+EEF with less significant Steppe added on. It also suggests an independent wave of Steppe went to all other parts of Europe. An Abstract earlier this year confirms CHG mixed with EEF in Central Turkey by 4500 BC, it'll be interesting to learn more about non-Steppe CHG. Results in CHG_K8 can't be taken too literally though.

Tuscan: 18% Potapovka + 32% CHG + 45% Iceman_MN + 4.99999999999999% WHG_CHG @ D = 0.0101
Tuscan: 16% Potapovka + 41% GeorgianCHG + 37% Iceman_MN + 5.99999999999999% WHG_CHG @ D = 0.0047
12% Yamnaya_Samara + 40% CHG + 37% Anatolia_Neolithic + 11% WHG_CHG @ D = 0.0134
10% Yamnaya_Samara + 50% GeorgianCHG + 29% Anatolia_Neolithic + 11% WHG_CHG @ D = 0.0055


35% OrcadianCHG + 27% CHG + 37% Iceman_MN + 1% WHG_CHG @ D = 0.0087
29% OrcadianCHG + 36% GeorgianCHG + 32% Iceman_MN + 3% WHG_CHG @ D = 0.0046


But the fits get worse when Turkish is used.
Tuscan: 9% Potapovka + 57% TurkishCHG + 29% Iceman_MN + 4.99999999999999% WHG_CHG @ D = 0.0507
Tuscan: 10% Potapovka + 59% Turkish_AydinCHG + 27% Iceman_MN + 3.99999999999999% WHG_CHG @ D = 0.0916
Tuscan: 11% Potapovka + 55% Turkish_KayseriCHG + 28% Iceman_MN + 5.99999999999999% WHG_CHG @ D = 0.0419
Greece: 5% Yamnaya_Samara + 69% TurkishCHG + 19% Anatolia_Neolithic + 6.99999999999999% WHG_CHG @ D = 0.0581

Fire Haired14
04-12-15, 04:53
However, what's the source of the additional CHG? It's explainable in various ways for southern Europe, but what about central and northwestern Europe?

The extra CHG is hard to explain. But anything is possible. It's totally possible groups from West Asia or Balkans rich in CHG migrated in North Europe during the Bronze age. The thing is Corded Ware, Unetice, and R1a-Z93 rich groups don't show the extra CHG. For the most part Bell Beaker doesn't either. Most Nordic LN/BAs do though. Corded Ware+Unetice lacking it suggests the extra CHG is not of Steppe origin.


In terms of the area of most interest to me, southern Europe, the results are also very interesting. As I've been saying ad nauseam, the "Caucasus" or "West Asian" in Italy can't be analyzed in a vacuum as something that only affected it and not neighboring countries. As I suspected, the CHG like levels are very similar in Greece, the Balkans, and Italy, and not all that much lower in Spain.

I've been saying for a while there's post-Neolithic West Asian ancestry in Italy. Are you convinced yet? I don't mean to be abrasive. The CHG/West Asian ancestry is significant. It mostly affected Greece-area and Italy, but we can also see it in Balkans, Iberia, Central Europe, and maybe all of Europe. It had to of involved a big movement of people across the Mediterranean sea. The Levant is certainly not the source. Northern West Asia, like Turkey looks like the source. I suspect it is the main source of J in Europe.

Also, something important to remember is there's a big genetic divide between Greece and Turkey. Greece has a lot more EEF and Turkey a lot more CHG. Turkey must have been overrun by CHG, while Greece got a watered down version of the same CHG migrations.


Missing, of course, are the figures for southern Italy and Sicily, which for some inexplicable reason were not provided.

He didn't include SouthWest Asia and North Africa either. I don't know why he didn't include South Italy+Sicily. For the other regions I think it's because Anatolia_Neolithic isn't a realistic ancestral population for them. They have ancestry that isn't CHG but it wasn't exactly Anatolia_Neolithic, so their scores might be weird. This test is focused on Europe, but also involves Caucasus+Turkey because they should be mostly CHG+EEF.

Angela
04-12-15, 06:15
Fire-Haired. I've been saying for a while there's post-Neolithic West Asian ancestry in Italy. Are you convinced yet? I don't mean to be abrasive. The CHG/West Asian ancestry is significant. It mostly affected Greece-area and Italy, but we can also see it in Balkans, Iberia, Central Europe, and maybe all of Europe. It had to of involved a big movement of people across the Mediterranean sea. The Levant is certainly not the source. Northern West Asia, like Turkey looks like the source. I suspect it is the main source of J in Europe.

I don't mean to be rude to you, Fire-Haired, but I have to confess that I am getting rather weary of you not remembering what I have said, or if you do remember it, not understanding it. Perhaps also part of the problem is that you are repeating the ideas of others without having thought it through for yourself. I really think you have absorbed a lot of information about these issues so I would strongly urge you not to do that.

What you have said repeatedly is that Italy has post Neolithic "West Asian" ancestry. What I have said repeatedly but have been unable to make you understand and what you still don't seem to understand even after looking at this calculator is that it is not just Italy that has West Asian ancestry. Please take a look once again at the CHG figures from the link you provided. It would also help if you would re-read my post.

The Greeks have more "CHG" ancestry (45%) than the Italians. The Bulgarians, Albanians and probably the rest of the Balkans have as much as Tuscany (40%). Goodness only knows what their levels were like before the Slavic migrations. The Northern Italians have less with 35%, the Slovenians 34%, the Poles 31%. I have been trying to get through to you that this is not just an issue that affects Italy.

Furthermore, unless you or your guru have a crystal ball there is no way of knowing when the CHG/West Asian started to enter the Balkans, Greece and Italy. For all we know some of it could have arrived in the late Neolithic. Oetzi already had some. Some could have come in the Copper Age. Have you heard of the Circum-Pontic Metallurgical Horizon? Some could also have come in the Bronze Age, as it did to other parts of Europe. Perhaps you have forgotten that Yamnaya was 50% CHG, and some of the CHG in this part of the world would have come with "Indo-Europeans"? Again, please re-read my post upthread. I doubt most of it in these areas came in the Iron Age because there is additional CHG even in northwestern Europe, and no Iron Age movement from the Near East would have reached northwestern Europe. The point is that we don't know, so we shouldn't be claiming that we do.


He didn't include SouthWest Asia and North Africa either. I don't know why he didn't include South Italy+Sicily. For the other regions I think it's because Anatolia_Neolithic isn't a realistic ancestral population for them. They have ancestry that isn't CHG but it wasn't exactly Anatolia_Neolithic, so their scores might be weird. This test is focused on Europe, but also involves Caucasus+Turkey because they should be mostly CHG+EEF.

Well, the Anatolia Neolithic is a very realistic ancestral population for southern Italy and Sicily and they should be included. They will turn out to be about 90% CHG and EEF.


There's little Steppe(based on EHG) in SouthEast Europe. Lots of West Asian CHG came in after Neolithic.
>South Europe is mostly Neolithic Euro. WHile nearby Turkey isn't. Turkey_Istanbul scores 23% Anatolia_Neolithic while Greek scores 35%.

There you go again. South Europe has from 30-40% Anatolian Neolithic. Even if you added in all the WHG, which you shouldn't, because some of it probably came with Indo-Europeans and historical area migrations, you wouldn't get over 50%.

There's also, in my opinion, no way to know at this point how much steppe based ancestry is in southern Europe. What if the steppe people who came to southern Europe were more heavily CHG than the Yamnaya people for whom we currently have samples. That would change all the calculations.

Danelaw
04-12-15, 15:53
Y-dna haplogroup J was already present in Anatolian Neolitich farmers who had no CHG. Most of the Caucasus stuff most likely was spread by females in the late neolitich-early copper age.

Alan
04-12-15, 16:37
The moment I saw this,

Yamna ~39% CHG, ~51% EHG and ~10% WHG.

in contrast to this where CHG is even slightly above EHG. The roles are exchanged it seems.

http://www.nature.com/ncomms/2015/151116/ncomms9912/images/ncomms9912-f1.jpg


I stopped bothering further. Honestly we could have placed bets early that something like this would happen. Name me only one single ancient sample so far which he hasn't changed it's admixture propotion in favor of "Northern" components in comparison to the original peer reviewed studies. Honestly someone must be blind to not realize that he has an obvious agenda and is crusading against "southern" admixture in anyways possible. As early as he gets his fingers on the ancient data suddenly a sample such as Andronovo turns from ~25% EF, 75% EHG/CHG into ~ 14% EF, 30% CHG, 34% EHG and 18% extra WHG.
So just for the future if you see any Eurogenes work, always add some 10-15% of Southern admixture and reduce the "Northern" admixture by the same amount and you have a perfect calculator :grin:

Danelaw
04-12-15, 17:15
According to that EHG had alreay a little bit of CHG and Amerindia. Moreover WHG, SHG and WHG score the same component, so it can't be right.

Fire Haired14
04-12-15, 17:26
@Alan,

Davidski doesn't have the bias you speak of. I see no indication at all. I only get tired of him making MA1, then EHG, then Yamnaya, then Sintashta/Corded ware in to ultra-macho guys who's Y DNA and languages(he's argued pre-PIE is from Siberia) always finds a way to dominate. And all southern admixture coming from Middle Eastern females. But at the same time he's aware of this, doesn't care, and doesn't allow it to make him create crazy theories.

Sile
04-12-15, 18:48
@Alan,

Davidski doesn't have the bias you speak of. I see no indication at all. I only get tired of him making MA1, then EHG, then Yamnaya, then Sintashta/Corded ware in to ultra-macho guys who's Y DNA and languages(he's argued pre-PIE is from Siberia) always finds a way to dominate. And all southern admixture coming from Middle Eastern females. But at the same time he's aware of this, doesn't care, and doesn't allow it to make him create crazy theories.

When he stops pushing everything into a eastern europe pattern with R1a then he will have better data............currently he tunnels everything into R1a as being everything for Europe.
I see he still has completely avoided all data from early neolithic central europe from EVERY paper in 2015 to use in his claculations.

MOESAN
05-12-15, 00:34
I have some doubts with Davidski sometimes, not for everyone of his works.
in<his late analysis he shows an increase in 'WHG' AND CHG %s in modern Europe (France, Slovenia) compared to EHG in the steppic ratio, so? But I agree he can be surprising...

MOESAN
05-12-15, 01:12
This analysis, if it stands up, raises some real questions about the estimates that have previously been proposed for the amount of actual Yamnaya ancestry in Europe, defined as a 50/50 split of EHG and CHG. It seemed to be the consensus on some sites that not only northern Europeans in general but also central Europeans are a 50/50 mix of MN and Yamnaya.I even saw people claiming the MN were almost totally wiped out and "replaced" by Yamnaya people. I don't see how that's possible if the percentages in this calculator are correct. The only people who might still work out as 50% Yamnaya might be the Balts or the Finns.*
This is an issue I've raised previously, saying that perhaps the group that went into central and northern Europe was more EHG shifted than Yamnaya, perhaps from the "Indo-Europeanization" of EHG forest steppe groups. However, the numbers here for a lot of even central European groups show less EHG and more WHG, and also more CHG. So, were there Indo-Europeanized groups which were more WHG shifted than the steppe people? ** However, what's the source of the additional CHG? It's explainable in various ways for southern Europe, but what about central and northwestern Europe?

Is it possible that the western Yamnaya region harbored a mix of WHG and EHG, or more WHG like EHG than those further east on the steppe, and somehow also harbored some CHG like population? I don't think that's impossible. When the J2 was found in Hungary I speculated as to whether a more Caucasus like population was starting to move into Europe at that time (mid to late Neolithic). I'd love to see the Kumtepe late Neolithic sample run through this calculator. The fact that the Greek sample from 4000 BC was very EEF like rather argues against that, but that might have been an anomalous result. Plus, I'd like to see that run through this calculator too. Or, the CHG might have begun to flow into Europe with the Copper Age. We have no idea of the autosomal composition of the CT people who moved onto the western steppe. Could they have mixed with a more western Yamnaya group?

In terms of the area of most interest to me, southern Europe, the results are also very interesting. As I've been saying ad nauseam, the "Caucasus" or "West Asian" in Italy can't be analyzed in a vacuum as something that only affected it and not neighboring countries. As I suspected, the CHG like levels are very similar in Greece, the Balkans, and Italy, and not all that much lower in Spain.
So, when did it arrive in Europe? Late Neolithic? Copper Age? Bronze Age? Or were there repeated flows?


Ed.

* In their case, there was probably also a reservoir of WHG or WHG/EHG hunters who became Indo-Europeanized.

One other thing that I thought I should mention in regard to CHG levels in southern Europe. All of the scenarios I mentioned above are possible, but there may be another factor as well. As I've speculated in the past, the "Indo-European" groups which went into the Balkans and Italy might have been CHG heavy, perhaps groups from the southern steppe. I've even proposed that it's possible that some of these "Indo-European" groups, specifically from the Yamnaya area from which we have samples, went around the Black Sea, through northern Anatolia and into Europe from that direction having picked up more CHG along the way, and so were more than 50% CHG by the time that they arrived. After all, the Yamnaya samples we have carried a form of R1b which is the most common R1b in southern Italy and Greece and the Balkans. Only samples from Anatolia and Greece from that period will tell us what happened.


- Some more HGs among steppic people arriving in Northern Europe? Very possible! (peoeple of old post-maglemose-swiderian cultures in the Forest Steppes?)
- SO, de facto, differences in components among Yamnaya and other steppic people? Very possible; already, metrically speaking, 2 sites of Khvalynsk showed big enough differences, as did the 2 Catacomb subdivisions (I ignore what kind of division helas! time? space?). A not-metric survey (discrete traits) about old Armenians and the diverses Steppes people (Ukraine for the most) of Chalco-Bronze showed sharings of some features, but a gradual enough divergence between sites. The samples we have of aDNA are scarce enough and concern some sites for every culture what is not sufficient to make states.
Even if of the same source, the Copper and then Bronze groups taking different ways would have had hard time to keep pure and unmodified, when we see the great instability/variability of their auDNA in all Europe at those times even when male lignages seemed unchanged!. and we are not sure Copper was introduced from ONE source only. The Copper in Andalusia, after BBs, saw people seemingly coming from Egea (Helladic) or surroundings; all the Eneolithic and Bronze in Iberia and France and Italy saw new physical southern subtypes (+ some 'dinarics') arriving and taking the strong side upon "old mediterraneans" of every kind (these last ones crossings with diverse Mesolithic types) EVEN in Greece which was not homogenous. Apparently, metrically, the most of first the change in South occurred around Copper which began sooner in South-East. a possible early osmosis between EEF and CHG could have begun in Anatolia before the metals but I think the most occurred at the transition period and after. through Anatolia or/and through Western shores of Black Sea (Romania-Bulgaria) the new groups had more chances to have an excess of CHG compared to EHG I think, whatever the culture I-E or not. So I agree for the most what you say.

Alan
05-12-15, 04:06
CHG K8 Spreadsheet (https://docs.google.com/spreadsheets/d/1_uKagAzyBnSzfI1rIEclIx6kfIE-NYpiWY5Bs0XqyAg/edit#gid=1521430867&vpid=A1)
CHG K8 by Region (https://docs.google.com/spreadsheets/d/1JELnEcj_Q8_AKtFK_4z4UxBSn37QLnvrpe8yTbmN_Jo/edit#gid=20013436&vpid=A10)

CHG_K8 is the newest ADMIXTURE test made by Daviski to estimate ancient ancestry. It represents the most up to date knowledge on European and to some extent West Asian genetics. What we've learned is Europeans are a mixture of 4 distinct populations in West Asia/Europe from 8,0000+ years ago who were distinct because of 1,000s of years of isolation. And this test IMO gives pretty accurate percentages of ancestry from each group.

Some here have miss interpreted and miss judged these ADMIXTURE tests made by Davidski. There's nothing to suggest he's using these tests for an agenda. He isn't even claiming they are actual percentages of ancestry from each ancient population. It's just a suggestion and trend according to him. ADMIXTURE method isn't perfect and no one ever claimed it is.

Summary of Information it Reveals.

New components absorb old components.
>EHG component can explain the high ANE scores in the Caucasus, Siberia, South Asia, and Americas.
>CHG can explain most ANE in West Asia.
>Roughly half of the WHG Europeans scored in ANE K8 is absorbed by EHG/Anatolia Neolithic. Actual percentage of ancestry from Western HGs like Loschbour isn't above 30% anywhere according to this test.

SouthEast Europe and Caucasus are mostly CHG+EEF
>Modern Turkey and Caucasus are mostly CHG.
>South Caucasus/Turkey is CHG+EEF. North Caucasus is CHG+EEF+EHG.
>Greece-area, Italy, and Iberia are mostly European Neolithic.
>Greece-area and Italy have a lot of CHG.
>There isn't much EHG and therefore Steppe in Greece-area and Italy. According to this test 10-20%(if you dive Yamnaya EHG by their EHG). CHG-related pop could be the main newcomers after Neolithic in these two regions.

North Europe is LNBA+CHG
>North Europe has significantly less EHG than LNBA North Europe. And has CHG unexplained by Steppe. Bronze age trade with South Europe?
>Balkans look like a mix of North Europe and Greece.

SouthWest Europe is CHG+EEF+LNBA
>Sardinians and Iberians have significant CHG.
>There's 20% Yamnaya in Iberia, but if you use a LNBA North European as source of Steppe, population turnover in Iberia is almost 50%. Looks like a lot of CHG also came into Iberia after Neolithic.



I have been saying this earlier and will repeat it, West Asia is not all EF+ CHG.

There is definitely at least one(possibly even two) population more, similar to EF and CHG, we haven't yet discovered. Those are the "South/Southeastern Farmers".


Northern West Asia is predominantly CHG+EF(~90-95%) with a little bit of "South farmer, Southeast Asian or East Asian" depending on the groups

I think some of it is getting eaten up by the "Afro_Eurasian component"


Southern West Asia (Arabia) is SF+EF+ CHG.


This might be a reason why North Africans and Arabians are not yet included in his calculator because they have ancestry which can not be explained by EF or CHG. And which I think is something Southern Farmer like which is basically a Afro_Asiatic version of EF.

Alan
05-12-15, 04:12
Y-dna haplogroup J was already present in Anatolian Neolitich farmers who had no CHG. Most of the Caucasus stuff most likely was spread by females in the late neolitich-early copper age.

Haplogroup J2a exclusively was present there, indicating that is was intrusive to the early farmers by CHG. Also it was percent-wise only ~6%. Even yDNA "I" was more with 12%

Alan
05-12-15, 04:18
There's also, in my opinion, no way to know at this point how much steppe based ancestry is in southern Europe. What if the steppe people who came to southern Europe were more heavily CHG than the Yamnaya people for whom we currently have samples. That would change all the calculations.


didn't just recently a paper appear which said that proto Latin scripts were found in Crete? All I can say that also fits with with Maciamos R1b root of Indo Europeans theory. That some of the Indo Europeans moved via Anatolia. I suspect that most of the Latin and Greek Indo European ancstors came via the Anatolian root. Explains also why most of the post Neolithic samples of Italy which were thought to present the first Indo European introduction into the region turned out predominantly EF. The first Latin speakers might indeed have been more heavily EF and CHG. Would also explain while everyhwere were there are Roman traces Haplogroup R1b and J2a are strong.

Danelaw
05-12-15, 09:54
Haplogroup J2a exclusively was present there, indicating that is was intrusive to the early farmers by CHG. Also it was percent-wise only ~6%. Even yDNA "I" was more with 12%

Anatolian Neolitichs score 0% of CHG, so you are dreaming. Virtually all J2a in Italy came clearly from Anatolian Neolitichs. Italians have the highest IBS affinity to Barcin farmers.

https://lh3.googleusercontent.com/xGIG-soMtsuZbxnxOQObQ0An5CcVvQk_pxKK2xt_ZWg=w200-h341-no

http://eurogenes.blogspot.it/2015/06/first-look-at-ancient-genome-from.html

In Italy there is about 30% of non Steppe derived CHG, in North Europe is 15-20%. Oetzi had 6% of CHG, so the total post neolitich non steppe derived CHG in Italy is 25%, which is not far from Hungarian levels (20%).


didn't just recently a paper appear which said that proto Latin scripts were found in Crete? All I can say that also fits with with Maciamos R1b root of Indo Europeans theory. That some of the Indo Europeans moved via Anatolia. I suspect that most of the Latin and Greek Indo European ancstors came via the Anatolian root. Explains also why most of the post Neolithic samples of Italy which were thought to present the first Indo European introduction into the region turned out predominantly EF. The first Latin speakers might indeed have been more heavily EF and CHG. Would also explain while everyhwere were there are Roman traces Haplogroup R1b and J2a are strong.

Greek and Latin are totally different languages. Italic languages are closer to Celtic ones, than Baltic languages are closer to Slavic ones. After that, Germanic languages are closest to the Italic ones.

Alan
05-12-15, 12:52
Have we another sock puppet here?

Danelaw
05-12-15, 13:34
Lack of arguments xD

Alan
05-12-15, 15:13
Nope it's just that I don't bother discussing with "no logic people".

yDNA J(J2a) in total among Anatolian farmers ~6%. Average CHG in Anatolian Farmers. There are some Anatolian farmer samples with 0 and other with ~5% CHG. going by this figures we don't even need to assume that there was an all male migration because 6% is low enough that selection for or against this component coult have appeared. But even without selection the few individuals with ~5% are enough to prove your claim that there is "no CHG in Anatolian farmers" wrong. What did you expect with 6% of J, that they end up as 30% CHG?

Why am I even discussing with this kid.

Danelaw
05-12-15, 15:50
Are you kidding me? Average score of CHG for Anatolian farmers (21 samples) is 0%.

https://docs.google.com/spreadsheets/d/1JELnEcj_Q8_AKtFK_4z4UxBSn37QLnvrpe8yTbmN_Jo/edit#gid=20013436&vpid=A10

Moreover the frequency of J2a1 among Barcin farmers was 7-8%. That's enough (with some founder effects) to explain the 10-15% of J2 in Italy.

http://www.biorxiv.org/highwire/filestream/7735/field_highwire_adjunct_files/0/016477-1.xlsx

Angela
05-12-15, 19:03
Maybe we're all looking at different data?

I just went to the link Fire-Haired posted in the OP. I see 21 Anatolian Neolithic samples. Five show some CHG: .01, .03, .04, .04, .04, and the rest are at 0. So, some CHG was filtering in, but not very much.

Someone refresh my recollection...these are only the earliest ANF samples? So, is that late Barcin sample included in this calculator?

More ancient samples will sort this out, and we haven't even seen the Lazaridis analysis of the Anatolian Neolithic (I even wonder if they'll wow us with some older, more southern samples as it's taking so long, or provide us with a comparison with Caucasus samples).

But, while waiting for that I think it would be interesting to have some other samples run through this calculator.

How about that Greek sample from 4000 BC which was said to be EEF like.

If it's true that there was no change in Greece in the Bronze Age, and the change was from the early Neolithic to the Late Neolithic, when exactly did things change and from where did the migrants come, and also what language did they speak? 4,000 BC seems pretty late to me, so it's important to be sure about the level of CHG in that sample. If there really is none, then did it arrive very shortly after that?

How about the J2 samples in Hungary? Have they been run through the calculator?


DaneLaw: Y-dna haplogroup J was already present in Anatolian Neolitich farmers who had no CHG. Most of the Caucasus stuff most likely was spread by females in the late neolitich-early copper age.

Well, actually, a few of them did, as I just pointed out. Also, if you're going to claim it was all or "mostly" spread by women from the Caucasus, to which specific mtDna subclades are you referring, and how did these women get to Europe? Did copper workers from the Caucasus bring them? If so, what yDna would they have carried? Or perhaps men from the Balkans also went all the way to the Caucasus to steal wives? What on earth were these Caucasus women like? :)


Danelaw: In Italy there is about 30% of non Steppe derived CHG, in North Europe is 15-20%. Oetzi had 6% of CHG, so the total post neolitich non steppe derived CHG in Italy is 25%, which is not far from Hungarian levels (20%).

Where are you getting this data? The link provided shows 35% Total CHG for Northern Italy, and 40% for Tuscany. (It's true that the Northern Italian percentage is almost exactly the same as that of Slovenia, and there are Central European and English samples with 34 and 36%. Heck, everybody in Europe gets at least 30%. The Greeks have even more, with 45%. The Albanians, Bulgarians, Bosnians, and a lot of the Romanians get 40% and higher.) There are no percentages for southern Italians.

How do you know how much of that is non-steppe? What is your source? It would be helpful if you would provide details on how you computed the percentages for both North Italians and Central Italians, and what the percentage of non-steppe CHG would be for the rest of Europe.


Originally Posted by Alan http://cdn.eupedia.com/forum/images/buttons/viewpost-right.png (http://www.eupedia.com/forum/showthread.php?p=472131#post472131) didn't just recently a paper appear which said that proto Latin scripts were found in Crete? All I can say that also fits with with Maciamos R1b root of Indo Europeans theory. That some of the Indo Europeans moved via Anatolia. I suspect that most of the Latin and Greek Indo European ancstors came via the Anatolian root. Explains also why most of the post Neolithic samples of Italy which were thought to present the first Indo European introduction into the region turned out predominantly EF. The first Latin speakers might indeed have been more heavily EF and CHG. Would also explain while everyhwere were there are Roman traces Haplogroup R1b and J2a are strong.


Latin is very close to Celtic. I would surmise that the speakers of the two languages originated in the same area. Later, Italo-Cetlic either split, or the groups were in proximity and the two language branches developed separated from a common root. Greek is on a different branch, and came with another migration, I think. I'm not sure about anything at this stage, but I think it's possible that some Indo-European speakers, perhaps CHG heavy and EHG light, entered Europe through Anatolia. That's the old "Greeks from the East" scenario. It's equally possible or perhaps more probable that Anthony is correct and that they came from the steppe and then down into the Balkans. If the latter is the case that leaves us with two scenarios? The CHG came before the "Greek Indo-European" speakers arrived, and the latter had little demographic impact, which might fit in with the paper that didn't see much change in the Bronze Age, or somehow the Greek Indo-European speakers came from an area on the steppe that was more CHG heavy or passed through some CHG heavy area.

Ed. Of course, this could just be a terrible calculator, which is pulling some EEF into CHG, inflating those levels, and pulling some EHG into the WHG cluster.

MOESAN
05-12-15, 23:14
Are you kidding me? Average score of CHG for Anatolian farmers (21 samples) is 0%.

https://docs.google.com/spreadsheets/d/1JELnEcj_Q8_AKtFK_4z4UxBSn37QLnvrpe8yTbmN_Jo/edit#gid=20013436&vpid=A10

Moreover the frequency of J2a1 among Barcin farmers was 7-8%. That's enough (with some founder effects) to explain the 10-15% of J2 in Italy.

http://www.biorxiv.org/highwire/filestream/7735/field_highwire_adjunct_files/0/016477-1.xlsx

is not this spreadsheet an Eurogene product? I find it strange: the three principal components of Europe seem very level with little variation between countries; I 've some doubt about this calculations...

Alan
06-12-15, 06:24
Looking at the spredsheet again. I get the sense that because EF and CHG are relatively close, some of the EF is getting eaten up by CHG. The same thing happens with some of the CHG that gets eaten up as EHG and some of the "WHG" ancestry in EHG gets eaten up by the WHG component. This might be the reason why the spredsheet has some flaws and this might be also the reason Satsurbila is not included. Something tells me if some CHG samples are included they will turn up again just as in "Kurds" calculator with ~20% EHG. The reason for that again could be that Eneolithic_Samara samples are taken as refference for EHG.

Alan
06-12-15, 06:52
Ed. Of course, this could just be a terrible calculator, which is pulling some EEF into CHG, inflating those levels, and pulling some EHG into the WHG cluster.


The calculator is not "terrible" but it has it's flaws.

And this is exactly what happens. Some of the EF is eaten up by CHG and some of the CHG is eaten up by EHG which imo causes this "extra WHG" effect in every of the ancient samples, which was never there in the original paper.

This is also why Sardinians end up with ~30% CHG ancestry in this calculator. At least ~20% of it is EF eaten up by CHG.

I expect that every West Eurasian population has at least ~40% more as their total EF is shown and 20% less CHG as their total numbers of this component.

Croatians taken as example they would have ~45/30/15/10 instead of 30/36/14/20 EF/CHG/EHG/WHG

I simply can't think of that most Europeans have more CHG than EF, never seen or heard that.

So that might be the reason why we have "extra WHG" popping up in Yamna which is actually simply EHG and some of the EHG is actually (~25%) CHG.

Yamna is like ~40% CHG and ~50% EHG in this calculator let's assume he took the Eneolithic_Samara samples with 25% CHG ancestry as refference.

So ~25% of the EHG wanders to CHG. makes ~ 12,5% extra CHG. Thats ~52.5% CHG. And since the "extra WHG" is basically EHG, that makes a total of ~47,5% EHG.

This are exactly the figures proposed by the peer reviewed paper (50/50).

So once again summarized.

-The EF figures are too low I assume every population to have more EF than shown here.

- This is because some EF is getting eaten up by CHG which is too high for most modern populations who should score on average ~25% less of this component.

-In return some of the CHG is getting eaten up by EHG which causes the effect that a WHG like portion of EHG ends up seperately shown as "extra WHG". I don't think that there is so much pre Indo European WHG left in Europe as shown here.

Danelaw
06-12-15, 10:35
Sardinians have 20% CHG, not 40. They have both more post neolitich basal eurasian and whg compared to Barcin farmer, and the software is using the CHG to compensate their souther pull.Satsurbila is too much low coverage to be used.

Angela
07-12-15, 20:56
Has anyone seen the spreadsheet of population averages? Without it there's no way of completely analyzing whether it's any good or not.